Nicolai Foss and Peter Klein have produced something rare: an even-handed account of a controversial trend in management. Typically, books on popular business topics fall into one of two equally unsatisfying types: either they are written by management gurus pitching trendy new philosophies, or by stuffy academics fighting reactionary battles against progress. Yet Foss and Klein have “managed” to avoid both extremes while also writing book that is accessible and engaging.

The topic of their book, the bossless company, has enjoyed a vogue for several decades among scholars and practitioners in management and organization, and, as Foss and Klein explain, has existed under various names and in various guises for at least two centuries. Today, it inspires experiments in business organization as well as enthusiasm among academics and in the business press. However, Foss and Klein are skeptical about the “bossless narrative”—the idea that the time of traditional hierarchy has come to an end, and that the future of business lies in flat organizational structures with little or no role for conventional managers. In their view, “the death of hierarchy has been greatly exaggerated” (Foss and Klein 2022, p. 14). To show why, and to highlight the value of hierarchy and managers, they use their book to explore the theory, history, and current practice of different kinds of bossless or boss-lite companies. In doing so, they keep the tone of their argument charitable and conciliatory: they are generous with their praise, and, like any good economists, take pains to examine the costs and the benefits of different bossless models. What matter most are the tradeoffs involved, which will be different for different companies as well as for the same company at different times.

Ultimately though, Foss and Klein are critical of the no-hierarchy model as a blueprint for most or all businesses. Their main criticism consists of two points. First, bossless companies often exaggerate the extent of their flat, decentralized structures—in practice, many such firms have very real bosses and hierarchies, although they may exist under different names or with less formality. Second, relatively bossless and flat companies are often successful for a limited time or only under highly specific circumstances: ultimately, they struggle with both scale and longevity. The upshot is simple: there is no one-size-fits-all type of organization, and anyone who says different is selling something.

Happily, Foss and Klein are constructive critics, and they have in mind an alternative to “flatland”. Rather than propose another flashy organizational model, their alternative is traditional management, albeit with qualifications. Their view is that while management may be mundane, it works. In particular, they defend the conventional wisdom that managers and hierarchies serve necessary and beneficial purposes in business organizations. However, a key point they also raise is that virtually all traditional hierarchies already have elements of delegation, flatness, flexibility, decentralization, and so on. As a result, defending traditional management does not imply supporting an inhumane authoritarian hierarchy, or some form of neo-Taylorism; a conventional hierarchical firm is not the dystopian bureaucracy suggested by critics and depicted so often in popular culture. Instead, it can be, and in practice often is, simply a framework created by entrepreneurs and top managers to aid in employee decision-making, particularly under dynamic, uncertain, and time-sensitive conditions. This approach is noteworthy because it means that Foss and Klein’s critique is not restricted only to the relatively small number of purely bossless companies. It applies also to many forms of organizing that partially adopt bossless or radically decentralized systems; for example, Google’s seemingly-abandoned (!) 80/20 rule. These are far more common than the extremes such as Valve, which is discussed below.

In fact, the subtitle of the book is almost a misnomer, as the perils of the bossless company are discussed mainly in the first part, while the second is devoted to explaining the importance, and indeed, the inevitability, of hierarchy. The rest of my discussion relates mainly to the first part, concerning the costs and benefits of organizing without bosses.

Leaky valves

Foss and Klein make effective use of the available evidence to support their critique of the bossless narrative. One of the more prominent examples of an allegedly bossless company that they discuss is Valve, a firm Ulrich Möller and Matthew McCaffrey have studied previously (Möller and McCaffrey 2021). Valve is a useful case study for several reasons, especially because it shows how, when formal hierarchy is abandoned, it is often replaced by informal hierarchy. Yet it is also a fascinating example of how organizational structure can profoundly influence the strategy–structure problem and the ways in which a company creates value.

Foss and Klein argue that one reason Valve’s model works is that its core business activities involve relatively few interdependencies, which in turn allows for a good deal of autonomy among employees and teams (Foss and Klein 2022, pp. 17, 110, 259–260). This description of Valve seems correct at the moment, but I believe the example that Foss and Klein use to illustrate it, namely, software development (particularly for video games), actually makes a different point than the one they intend. As others have argued, modern game design is a tangle of complex and interdependent tasks that require considerable coordination, usually under tight and highly public deadlines, with the whole process being fraught with uncertainty (Möller and McCaffrey 2021, pp. 227–229). If this is the case though, then according to Foss and Klein’s argument, Valve’s bossless model should have run into financial trouble a long time ago. So how did it overcome these problems? The answer is simple: it stopped making games.

This point is worth emphasizing because it highlights Valve’s commitment to flatness, and its corresponding approach to strategy and structure. In most of the cases discussed by Foss and Klein, when a bossless model gets in the way of the financial health of the company, it is abandoned or at least revised. Yet instead of abandoning flatness to protect its future, Valve changed its future to protect its flatness. In practice, this has meant many failed and abandoned projects resulting in a general move away from software development and toward more modular and less interdependent projects. Even more significantly, this change in the company’s strategy was not decided democratically by employees, but has usually been credited to the decisions of founder and major shareholder Gabe Newell (the true entrepreneur in the company), who has repeatedly stated that he is encouraging teams to pursue non-game projects, especially experiments with new hardware, most recently the Steam Deck. And although hardware development has its share of complexities too, in some ways it also has more easily definable goals than software in terms of a clear vision for the final product.

For all intents and purposes, Valve is not currently a software company. In fact, if anything, at the moment, Valve does not seem to know what kind of company it is, just that it's a bossless one. It has certainly struggled with its own organizational model and sense of direction as a firm though. This point has been raised explicitly by journalist Geoff Keighley, who has provided the most recent and reliable update about the ‘state of the disunion’ at Valve. Keighley has embedded himself in the company several times in the past during the crucial “final hours” of production on some of its most anticipated games. Most recently, he was granted full access to the company as it finished and published Half-Life: Alyx in 2020. Based on extensive interviews with employees, Keighley wrote a short history of the company’s struggles with self-organization that explicitly raises many of the issues described by Foss and Klein. Keighley even repeatedly draws on Adam Smith’s metaphor of the invisible hand to explain how “The company’s self-professed, no-hierarchy organizational structure has led to inefficiencies and periods of creative drought” (Keighley 2020, Introduction).

Several of Keighley’s discoveries are worth mentioning. For example, like other flat companies such as Oticon, Valve struggled with having too many new ideas and not enough centralized management to help sort through them and allocate scarce company resources effectively: “The invisible, self-directed hand of Valve… led the company in a dizzying array of directions” (Keighley 2020, Chapter 3). This led to creative inertia and eventually to the collapse of most efforts to develop new games.

Valve’s employees were keenly aware of the problem, but without managers to channel their creative energies, project after project floundered. By 2013, employees had begun referring to the creative state of the company as, “The Wilderness” (Keighley 2020, Chapter 3). As Keighley explains,

Most of the time, your job at Valve is to make great games. But sometimes, since there is no real management layer, employees have to take on the “people stuff” to steer the invisible hand of Valve in a particular direction. By early 2016, it was becoming clear to [employees] that an intervention of sorts was needed to help re-spark Valve’s creative output. (Keighley 2020, Chapter 5)

Although Valve has not formally abandoned its bossless model, some management was required in order to get Half-Life: Alyx off the ground, including coordinated discussions about company culture. Furthermore, Alyx only became a viable project because it was developed to complement Valve’s experiment with Virtual Reality hardware: this enabled the project to become attractive to a large enough number of employees that it was able to overcome the resistance of Valve’s internal, informal hierarchy to making new games. The process by which employees overcame this inertia was even described by one employee as, “the story of how we fixed Valve” (quoted in Keighley 2020, Chapter 1). Again, there is no evidence of any formal or lasting change in the organizational design of the company, but in practice, it was circumvented to produce Alyx.

Ultimately, when faced with the limitations of its own lack of organization, Valve chose to pursue different kinds of value creation rather than to reorganize. The same kind of choice is evident in Valve’s lack of traditional departments, especially a dedicated marketing team and regular customer support (the latter is typically outsourced to the Steam community). This approach works for Valve in the sense that it conserves company resources. However, it has also had a negative effect on customer relations, and in fact, Valve is legendary for its hands-off approach in this area. To take only one example, players of Valve’s classic Team Fortress 2 have recently abandoned the game in droves after it became essentially unplayable due to the proliferation of automated “bot” accounts. Valve largely ignored the problem until a social media campaign in 2022 convinced it to patch the game more fully.Footnote 1 Up to that point though, the game had only been updated sporadically, and despite its consistent popularity, seems not to have had a large team supporting it, to the outrage of some of its most loyal customers. We can summarize Valve’s approach to these kinds of problems using Herbert Simon’s parable of the two watchmakers who keep getting distracted by customer phone calls (Foss and Klein 2022, pp. 167–168). When confronted with the distraction, Valve’s approach is simply to cut the phone line!

The Valve example supports Foss and Klein’s arguments in that it shows flat models are not generalizable: most companies contain relatively specialized human and physical resources, and few indeed have the luxury of choosing to produce completely different products or services in order to preserve the organization’s design. And of course, most companies are not free to ignore their most passionate customers and keep them in the dark about their plans for new products or support. However, Valve does have these luxuries, mostly due to the financial success of Steam.

Challenges and opportunities for research on the bossless firm

An underlying theme of this book is that, despite the popularity or trendiness of flat, bossless companies, there is still surprisingly little research about them. It is therefore worthwhile to suggest some ideas about different kinds of research that will be valuable in the future.

Foss and Klein conclude Part I of their book, the more critical part, with a discussion of the current data on bossless firms, or lack thereof. As they explain, there is little systematic evidence that makes the case for the success and for the generalizability of bossless models (Foss and Klein 2022, pp. 117–118). They mention the lack of industry- and national-level data sources, but it is worth pointing out that the nature of the bossless company itself makes data gathering far more difficult than in traditional firms (see also Lee and Edmondson 2017). In a traditional firm, an established chain of command results in relatively well-defined rules and decisions that can be documented and analyzed. However, in a flat company where there is no such chain of command, and in which communication is largely informal with little or no reporting to managers, the most basic data points often do not exist, or cannot be effectively documented. A good example that is raised throughout the book involves informal hierarchy: bossless firms are prone to replacing formal authority with office politics. Enough current or former employees at companies like Valve (even those with good incentives to say otherwise) have described this problem that we know it exists. However, just how pervasive it is—and what its long-term effects on the health of the company are—is unknown. This then is one vital topic for future work.

Unfortunately, the lack of reliable data hinders analysis on both sides of this debate, and researchers are often not in a strong enough position for us to pass judgment either way. To take one small example, when Foss and Klein discuss Zappos’ reputation for exceptionally supportive customer service, they suggest that the anecdotes behind this reputation may be exaggerated (Foss and Klein 2022, pp. 80–81). Yet while the extreme stories about outstanding support are just that, extremes—it is simply impossible, for instance, that employees could regularly spend ten hours on one phone call with the same customer)—there does not appear to be any evidence that these stories are “too good to be true” to an extent that would undermine the company’s narrative about its organization. Again, we simply need more reliable data.

However, somewhat paradoxically, we also need more theorizing of the bossless company (see also Lee and Edmondson 2017). After all, we cannot know what to look for empirically without a consistent, causal, realist account of how such firms emerge, develop, and ultimately, succeed or fail. Fortunately, some recent work is already showing promising results along these lines. For example, Raveendran et al. (2022, p. 825) develop an agent-based model of employee self-selection to explain the “trade-off between intertemporal coordination failure… in traditional staffing [versus] interpersonal coordination failure… in self-selection.” They find that more manager-less companies tend to face the latter problem, which leads to over- and understaffing of tasks. Similarly, Ketkar and Workiewicz (2022) formally model self-selection, finding that it works best when human capital is scarce relative to the number of opportunities the organization faces. Both results are illustrated once again by Valve, which seems to struggle with matching team size to project needs. In this and in other cases of bossless organizing, self-selection is often carefully channeled through rules that govern how it takes place (Möller and McCaffrey 2021; Ketkar and Workiewicz 2022). Crucially though, these rules have to be designed and enforced by someone in a hierarchy. There’s just no getting away from managers; hence, Foss and Klein’s book. Foss and Klein also mention technological change as a key driver of efforts to delayer and decentralize business, and there are undoubtedly many other factors that encourage or discourage the development of flatter firms, and that influence which specific approach that a company (or rather, a CEO or an entrepreneur) chooses to adopt. In particular, we need some essential conceptual work on the institutional foundations of the bossless company, as well as the relationship between these firms and public policy.

To this end, Foss and Klein deserve credit for emphasizing that in good research, novelty is not an end in itself. There is plenty of wisdom in the classics of economics and management that has stood the test of time and on which we should continue to draw for insight into contemporary problems (Foss and Klein 2022, pp. 271–272; Puranam et al. 2014). In fact, a more careful reading of the classics might have reminded us of the limitations of certain fads in popular management. Applying the Whig theory of history to management and organization can only encourage superficial trends and result in the loss of knowledge.